© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 9492
    Prev
    Next
    INDEX
    Explanations

    passages discussing controversial or impactful events or situations

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    684
    +0.12
    0.6%
    399
    +0.12
    0.6%
    496
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    399
    +0.12
    0.02
    324
    +0.12
    0.03
    380
    +0.10
    0.02
    Negative Logits
    gerald
    -0.67
    ynski
    -0.67
     additions
    -0.64
    hesda
    -0.64
     Expend
    -0.63
    ģĸ
    -0.62
    Missing
    -0.61
    ucket
    -0.61
    ãĥł
    -0.60
     [&
    -0.59
    POSITIVE LOGITS
     aloud
    1.10
     louder
    1.08
     loudly
    0.97
     talk
    0.93
     tongues
    0.92
     voic
    0.92
     loud
    0.91
     speeches
    0.90
     language
    0.88
     Mandarin
    0.84
    Activations Density 0.196%

    No Known Activations